\documentclass[12pt,a4paper,notitlepage]{article}
\usepackage{setspace} % Allows me to choose the %\doublespacing command
\usepackage{fullpage}
\usepackage{amsmath}
\usepackage{amssymb}
	\newenvironment{example}[1][Example]{\begin{trivlist}
	\item[\hskip \labelsep {\bfseries #1}]}{\end{trivlist}}
\usepackage{hanging} % This is key for making the citations look good
\usepackage{url}
\usepackage[pdftex]{graphicx}
\newcommand{\tab}{\hspace*{2em}}%allows using the \tab command
\usepackage[margin=2.5cm]{geometry}
\usepackage{tikz}
\usepackage{lscape}
\usepackage{harvard}
\usepackage{multicol}
\usepackage{verbatim}
\usepackage{standalone}
\usepackage{caption}
\usepackage[normalem]{ulem}
\setlength{\captionmargin}{30pt}
\definecolor{navy}{RGB}{0,0,175}
\definecolor{maroon}{RGB}{175,0,30}
\usepackage{setspace}
\usepackage{footmisc}
\usepackage{appendix}
\usepackage{titlesec}
\usepackage{titling}
\usepackage{rotating}
\usepackage{booktabs}
\usepackage{array}
\newcolumntype{L}[1]{>{\raggedright\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
\newcolumntype{C}[1]{>{\centering\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
\newcolumntype{R}[1]{>{\raggedleft\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
\usepackage{tgtermes}
\renewcommand{\thetable}{\Roman{table}} % Makes tables have Roman numerals
%\usepackage[none]{hyphenat}

\tolerance=1
\emergencystretch=\maxdimen
\hyphenpenalty=10000
\hbadness=10000

\renewcommand\footnotelayout{\,\doublespacing\fontsize{12}{12}\selectfont }
\setlength{\footnotesep}{\baselineskip}

\begin{document}

\title{\textbf{Vote Buying by the United States \\ in the United Nations}\thanks{We would like to thank Josh Clinton, Matthew DiLorenzo, Nick Eubank, Adriane Fresh, Will Howell, Brenton Kenkel, Trish Kirkland, and seminar participants at the University of Chicago, Vanderbilt University, and the 2017 annual meeting of the Midwest Political Science Association, three anonymous reviewers, and the editorial staff for their helpful assistance.}}
\author{
Dan Alexander \\ {\small \emph{Vanderbilt University}}
 \\ {\small dan.alexander@vanderbilt.edu} \and Bryan Rooney \\ {\small \emph{Carlos III-Juan March Institute of Social Sciences}} \\ {\small brooney@clio.uc3m.es}}
\date{\small\today}

\maketitle
\thispagestyle{empty}

\begin{abstract}

A compelling body of scholarship establishes a robust link between a state's election to a rotating membership on the United Nations Security Council (UNSC) and increases in foreign aid that state receives, especially from the United States \cite{Kuziemko2006,Vreeland2014}. This literature suggests this provision of aid reflects attempts to procure the support for US positions on UNSC votes but does not provide causal tests of this claim, a claim that has important implications for the legitimacy of UNSC decisions. This paper seeks to determine whether the US distributes foreign aid to ``buy votes.'' We generate theoretically-motivated hypotheses about the relationship between relative voting congruence with the US and the receipt of US foreign aid. Leveraging natural variation from the rotating structure of non-permanent UNSC members, we uncover a causal relationship consistent with the claim that the US uses foreign aid to procure support for its positions on the UNSC.

%150 words

\end{abstract}

Keywords: United Nations Security Council, foreign aid, vote buying

\begin{centering}

Word Count: 8492

\end{centering}

\thispagestyle{empty}

\clearpage %This starts a new page

\doublespacing
%\onehalfspacing

\setcounter{page}{1}

\section*{Introduction}

Scholars have demonstrated in great depth that non-permanent members of the United Nations Security Council (UNSC) systematically receive enhanced financial assistance while serving on the council \cite{Kuziemko2006,Vreeland2014}. Consistent with expectations about aid's strategic nature \cite{Alesina2000}, states with a vested interest in the outcome of Security Council resolutions, primarily the United States, contribute a great deal of this financial aid. Many of these studies further claim that this distribution of foreign aid by the United States appears to be an attempt to buy influence on the Security Council, a claim with significant normative implications for the legitimacy of UNSC resolutions.\footnote{Questions about vote buying in the United Nations remain as relevant as ever. For example, President Trump threatened to cut aid to those countries that voted for a resolution condemning US recognition of Jerusalem as the capital of Israel \cite{landler2017}.} The evidence these studies present to support the claim that the US is attempting to buy votes, however, is by the authors' own admission less conclusive than the evidence that rotating members receive a bump in foreign aid for their two-year terms. We seek to fill this gap in the literature by developing a research design and measurement strategies that allow us to determine more definitively whether the US distributes foreign aid in a manner consistent with vote buying.
%We review the suggestive evidence presented in prior work as well as the limitations those studies discuss.
%\cite{Kuziemko2006,Dreher2009,dreher2009global,Vreeland2014}.

We look to the theoretical literature on vote buying to distill testable predictions about the optimal (i.e., most cost-effective) distribution of payments for an entity seeking to buy influence in a deliberative body such as the UNSC. Specifically, theory predicts that payments from the US increase as the propensity of a state to vote against the US rises, until the desired level of support is reached. We obtain causal identification by leveraging natural variation in relative voting congruence with the US provided by the staggered rotation of non-permanent members on and off of the Security Council. We find a significant and robust relationship between relative propensity to vote against the US and the relative receipt of US economic and military aid for those states we predict the US will target to secure the necessary votes for passage of a resolution. Further, our evidence suggests that the deployment of aid does not extend to attempts to secure unanimity - that is, to those members whose voting is least congruent with the US and whose votes are not strictly necessary for passage.
%This prediction aligns with previous research, while also being sharper in ways that make it more amenable to empirical validation.

Finding that the US directs foreign aid in a manner so consistent with the predictions of vote buying provides the strongest evidence to date that US outlays of foreign aid to temporary UNSC members represent attempts to gain influence over UNSC decisions. This confirms the claims of earlier work while simultaneously validating vote buying theory's predictions in a new context. Perhaps most importantly, these findings underscore the questions previous work has raised about the meaning of UNSC decisions. While approval by the UNSC confers legitimacy on the actions of states in both the international \cite{hurd2002legitimacy} and domestic arenas \cite{chapman2004united}, its decisions may not reflect the collective will of the international community.

\section*{The United Nations Security Council and Foreign Aid}

The United Nations Security Council is the principal UN apparatus charged with the maintenance of international peace and security. If the Security Council determines that there is a threat to global peace or that an act of aggression has occurred, it determines what measures -- up to and including the use of military force -- are necessary to create or restore peace. Under the Charter of the United Nations, the UNSC has the authority to enact binding resolutions and require that all members of the UN carry out its decisions. The Security Council thus has substantial power on questions of great importance in both the international and domestic political arenas.\footnote{In 2017, the UNSC passed resolutions relating to refugee crises in Africa, the ongoing conflict in the Middle East, and threat of terrorism. The agenda is largely driven by emerging or escalating crises and threats to international security, consistent with its mandate.}%***

Since 1965, the Security Council has consisted of 15 member states. Of these, five of the member states -- the United States, the United Kingdom, France, Russia, and China -- remain permanently on the UNSC, each having the power to veto any substantive resolution brought before the council. In addition, there are 10 non-permanent members that hold their seats on a rotating basis. Each temporary member serves a staggered two-year term, starting on January 1, after having been elected during the previous year. Elections to the Security Council occur by geographical region. While some argue that states regularly seek out and campaign for a position on the Security Council \cite{malone2000eyes}, often a fairly strict norm of rotation governs selection,\footnote{Politicking to become a member of the UNSC, to the extent that it exists, likewise occurs regionally \cite{Vreeland2014}.} and there are few consistent political or economic predictors of election to the UNSC \cite{bueno2010pernicious}.

More interesting is what occurs after election to the UNSC. Recent studies have convincingly demonstrated that UN member nations receive more foreign aid and financial assistance during a rotating membership on the UNSC. A non-permanent member of the council experiences an 8\% increase in development aid from the United Nations \cite{Kuziemko2006}. Aid from major powers similarly spikes when countries rotate onto the UNSC \cite{Vreeland2014}. The number of World Bank projects a country receives increases by roughly 10\% when the state is a temporary member of the UNSC \cite{Dreher2009}. Additionally, temporary UNSC members are more likely to receive IMF loans and face fewer conditions on said loans while on the UNSC than when they have rotated off the council \cite{dreher2009global}.

The direct and indirect roles of the US in much of this increased assistance has led scholars to ask whether the US provides foreign aid to temporary members of the UNSC in an attempt to buy votes. Vote buying in this context refers to the offer of a payment that is in some way contingent on vote choice \cite{nichter2008vote}. The United States, as the predominant security actor in world politics, has a sustained interest in many UNSC decisions, and support from the UN can ease the burden of US action both militarily and financially, providing a rationale for buying votes. Further, since important security questions regarding any threats to international peace come before the UNSC, the Security Council acts as a signal of the legitimacy of any international security action \cite{hurd2002legitimacy}. Legitimacy is important both internationally \cite{Claude1966} and domestically \cite{Finnemore1998}, where UNSC decisions impact the level of domestic support for US military action \cite{chapman2004united,Grieco2011}.

\citeasnoun{Voeten2001} suggests that the United States uses its considerable might to influence Security Council decisions, while \citeasnoun{dreher2008does} and \citeasnoun{carter2015democracy} present evidence that the US impacts certain UN General Assembly votes successfully through both threats and enticements regarding aid. \citeasnoun{Kuziemko2006} find that a non-permanent member of the Security Council experiences a 59\% increase in total aid from the United States during its two-year term, with larger increases in years the authors classify as important.
%They favor the interpretation that this represents vote buying, especially as aid increases in years the authors classify as important. They note, however, that the inability to observe how the state would have voted in the absence of attempts at vote buying plagues any study of vote buying in the UNSC (p. 924).
The case that the US is buying influence on the Security Council is made most forcefully by \citeasnoun{Vreeland2014}. In this study, the authors examine the disbursement of foreign aid to UNSC members from a number of countries and international organizations. They offer compelling qualitative evidence that the practice of buying votes on the UNSC is common, at times even explicit. In one such instance, the US cut all of its \$70 million in aid to Yemen because of a ``no'' vote on the resolution authorizing Operation Desert Storm (67-69). Empirically, however, while they find that neither allies nor enemies drive the result of increased bilateral aid from the US for UNSC members, they are unable to demonstrate that such countries receive significantly less additional aid from the US than ``swing voters'' (175-181).
%The authors' most direct test of their thesis examines whether the US targets ``swing voters'' on the Security Council (pp.\;175-181). They argue that a vote buyer need not pay its closest friends and would not pay its enemies.

In explaining the mixed support for their thesis, \citename{Vreeland2014} remark on a number of difficulties in carrying out the analysis, noting at the outset that it ``demands a lot from the data'' (p.\;175). They concede that their measures of political affinity may be too blunt, suggesting the hypothesis itself may not have been sufficiently detailed to support rigorous testing (p.\;176). Further, in \citename{Vreeland2014}'s analysis, as well as the analysis  from \citename{Kuziemko2006} on which it is based, the authors limit themselves to variation in the states elected to the UNSC, rather than exploiting the additional, significant variation created by the rotating on and off of half of the council during each state's term.

Prior work has mustered substantial quantitative and qualitative evidence that the allocation of aid constitutes vote buying. Yet the authors of these studies recognize the limitations of their analyses and helpfully highlight these issues for those conducting future investigations. Following their cues, in the following section, we isolate findings from the theoretical literature on vote buying, consider the testable implications for this theory in the context of the UNSC, and develop a considerably sharper hypothesis about the pattern of bilateral aid we would expect the US to display if it is, in fact, buying votes on the UNSC.

\section*{Predictions from Vote Buying Theories}

%Scholars have examined vote buying by party organizations or other interest groups both of individuals in elections \cite{stokes2005perverse} and of members of a legislative body \cite{groseclose1996buying}. For example, students of the US Congress have examined the use of distributive outlays as a means to procure support for votes \cite{Alexander2016} and the exchange of a vote on one issue in exchange for a vote on another issue, a process commonly known as log-rolling \cite{stratmann1992effects}.

When investigating vote buying within the UNSC, we consider not the buying of specific votes or outcomes, but rather affinity. Because of the timing between foreign aid appropriations and individual votes on the UN Security Council, states are likely unable to redirect aid quickly to reflect the outcome of a single vote, and unwilling to undertake such a large bureaucratic effort outside of extreme circumstances. Reactions as swift and pointed as the US response to Yemen in the early 1990s constitute rare events, even if the underlying tendency is present. Instead, states set up flows of foreign aid to temporary members to ensure a pattern of aligned voting during their time on the Security Council. It is instructive to ask how the United States would most cost effectively allocate this aid to members of the United Nations Security Council to procure support. Accordingly, we turn to the theoretical literature on vote buying in legislatures, which takes as its motivation the question of how to cost-effectively trade payments for votes.

Models of both single \cite{SnyderJr1991} and competing \cite{Dekel2009} vote buyers in legislatures generate the same prediction regarding the deployment of aid in a deliberative body. These theories take as their premise a vote buyer lacking sufficient support for a proposal she prefers to the status quo. In both the competitive and non-competitive setting, if the vote buyer purchases any votes at all, she will begin with the member that requires the smallest payment to change her vote; she will proceed to buy the votes of increasingly costly members until her proposal has the required level of support, with the relative price paid increasing in the member's disagreement.\footnote{These theories do not model endogenous agenda formation in the context of vote buying, but it would only alter the frequency -- not the pattern -- of vote buying. Further, while a state may exaggerate the amount it requires to change its vote in accordance with US preferences, it could only do so in as far as the US could not buy the next most costly state as a substitute. A utility-maximizing state would not forgo a price at or above the value it places on a sincere vote. The literature has not explored these market dynamics in full, and it is beyond the scope of this paper, but the predictions isolated here nonetheless represent an ``approximate equilibrium.''}

Applied to US deployment of foreign aid across members of the UNSC, these theories predict payments will increase as the initial propensity to vote against the US rises. Of course, no vote buyer wishes to purchase more support than necessary. Passage of UNSC resolutions requires nine affirmative votes out of the fifteen members, with no vetoes from permanent members. It has often been the case that Russia and/or China will abstain on an individual resolution, thus requiring up to six votes from rotating members. If the US seeks only a minimal winning coalition, it would procure the six least costly votes. UN scholars, however, emphasize the empirical regularity of unanimity on the Security Council as well as its role in bestowing legitimacy \cite{dryzek2006reconciling}. If the US desires unanimity, we would predict that aid increases as propensity to vote against the US increases across all ten non-permanent members. We test both of these possibilities below.
% Throughout, the null hypothesis is that there does not exist a positive relationship between propensity to vote against the US and US aid to the non-permanent members of the UNSC.

\section*{Research Design}

Examining the relationship between the propensity to vote against the US and foreign aid received presents both design and measurement hurdles. Specifically, we must construct a measure of propensity to vote against the US that is not reflective of Security Council behavior potentially related to US aid, isolate exogenous variation in a state's relative propensity to vote against the US, and account for confounding patterns potentially present in cross-sectional data. We begin by clarifying our explanatory variable of interest, which we construct from an estimate of states' propensities to vote against the US. We then specify our identification strategy, which exploits the staggered rotating structure of the non-permanent members of the UNSC, along with fixed effects. Finally, we discuss the outcome measures of interest and controls.

We have argued that evidence supportive of vote buying would entail increases in US foreign aid as propensity to vote against the US increases and thus the aid necessary to change a state's vote increases. An assumption underlying this prediction is that the lower a state's propensity to vote alongside the US on the Security Council, the more aid necessary to change its vote, all else equal. In vote buying theories, we may simply order members of the deliberative body by the size of the payment they would require to change their vote. In the real world, we must take ideological misalignment as a proxy for the costliness of a member's vote. Propensity to vote opposite the buyer need not be related one-to-one with the costliness of the vote, only correlated, to serve as a proxy for costliness, and thus we take propensity to vote against the US as our quantity of interest.\footnote{To the extent that a state's propensity to vote against the US is an imperfect gauge of the aid a state requires to change its vote, the measure would introduce noise into the analysis. We discuss one way to reduce such noise in our robustness checks.} %***

To obtain a measure of a state's propensity to vote against the US, we must first construct a measure of foreign policy similarity with the United States for each nation. We generate a yearly estimate of the probability that a state votes the same way as the US using voting in the General Assembly (UNGA) \cite{Voeten2009data}. Scholars have regarded voting patterns in the UNGA as evidence of the similarity of strategic interests between states both historically \cite{alker1964dimensions,dixon1981emerging} and more recently \cite{Kim1996,Gartzke1998Kant,voeten2004}, particularly when studying the provision of aid \cite{Alesina2000,Vreeland2014}. Importantly, although there is some evidence of vote buying in the UNGA, this evidence is mixed in regards to the conditions under which states can be bought \cite{lai2006impact,carter2015democracy}, and few votes appear to be affected \cite{wang1999us}. Further, despite strategic selection at play in the agenda setting of the UNGA, states can nevertheless demonstrate a wide variety of preferences on the issues considered in the body. In 2016 alone, the UN drafted and voted on resolutions regarding refugees from Georgia, violations of international law on human rights in Syria, and entrepreneurship for sustainable development \cite{unorg2017}.
%\footnote{Measures of states' voting congruence based on military and non-military subsets of votes (based on coding by \citename{Voeten2009data}) were highly correlated with the measure based on all votes as well as with one another. Since in some years these measures were based on very few votes, and since we have no strong theoretical reason to favor one cut of UNGA votes over another, we proceed with a measure of voting congruence with the US that is based on all UNGA votes in a given year.}

We employ UNGA voting from the session before a member's term on the UNSC to measure states' propensity to vote against the US, fixed as the state's propensity to vote against the US for the two years of its membership on the UNSC. The decision to use and hold constant a measure of states' relative agreement with the US from before their UNSC terms begin serves three purposes. First, realizing that vote buying in the UNSC may affect voting in the UNGA, this approach isolates the UNGA-based measure from vote buying that may be occurring on the UNSC, while still approximating a state's inclination to vote with the US. Second, while a state's voting inclinations may not be entirely fixed on a year-to-year basis, there is evidence that voting within the UN is relatively stable \cite{holcombe1996stability}. Further, in each model we control for systematic factors that may dramatically alter a state's voting inclinations during their time on the council, such government turnover, which would call into question the assumption of a fixed inclination to vote with the US. Finally, the use of a fixed score is fundamental to our identification strategy. We ultimately wish to use a relative measure of states' propensity to disagree with the US, since a change in a rotating member's relative propensity within its UNSC term that resulted from actions it took would raise concerns that the state's relative propensity to vote with the US is endogenous to its receipt of foreign aid.

The estimate itself follows the approach of \citeasnoun{Fowler2016}.\footnote{A summary of the construction of this variable may be found in Appendix \ref{sec:const}.} Specifically, for each year in our sample and across all pairs of states and UNGA resolutions on which the US cast a `no' or `yes' vote, we formulate a variable taking the value of 0 if the state voted the same way as the US, 1 if the state voted in opposition to the US, and 0.5 if the state abstained. We regress this variable on state fixed effects, suppressing the constant. The estimated state fixed effects for each year serve as our measure of each state's propensity to vote against the US.

Using each state's individual estimates, we calculate a rotating member's relative propensity to vote against the US by dividing it by the sum of all rotating members' propensities to vote against the US in that year. We call this a rotating member's \textit{Contribution to Disagreement} as it represents a state's share of the overall propensity of all rotating members to disagree with the US on votes in a given year. This captures how unlikely a state is to vote alongside the US without enticement, relative to the other temporary members on the UNSC.%***

Examining vote buying in this context requires a careful identification strategy. The staggered rotation structure of the UNSC provides a unique source of exogenous variation in relative propensity to vote against the US. Each rotating member serves its entire term with four other rotating members, and serves each half of its term with a different, additional set of five temporary members. Having held fixed each state's propensity to vote against the US for their two-year UNSC term, a member's \textit{Contribution to Disagreement} will shift within their term only from this replacement of five of the temporary members from year to year.\footnote{Shifts in a state's voting inclinations would therefore not threaten the identification strategy as they are not systematically related to the countries rotating on and off during a member's term.}

Employing only this source of variation in relative propensity to vote against the US constitutes a fairly conservative test. First, we have only one such rotation to exploit for each member, i.e., we only get one shot to observe a change in relative propensity to vote against the US. Second, the new member will be from the same region of the world as the state it replaces, and if there is greater correlation in the ideologies of successive members from the same region than across regions, dramatic changes to the ideological landscape of the UNSC would be unlikely. Empirically, just 20\% of the variance in the \textit{Contribution to Disagreement} measure is variation from within each state's term.

Simply performing a cross-state analysis would risk confounding evidence of vote buying with other systematic patterns between states' voting tendencies and the amount of aid they receive. For instance, if poorer states vote in opposition to the US, then their higher receipt of aid would inflate evidence of vote buying. With the addition of state-specific effects and year dummy variables, as well as our battery of controls, we leverage within-state changes in propensity to vote against the US to examine the relationship between propensity to vote against the US and aid across states.\footnote{This also helps mitigate the noise introduced by the imperfect correlation of a state's propensity to vote against the US and the payment it would require to change its vote.}

As stated, we are looking for evidence of vote buying in the form of proportional changes in aid. If the US buys a given country's vote in both years, we assume it must pay the same amount, since the state's absolute opposition to the US has not changed. Rather than a change in the absolute amount a country receives, we are looking to uncover the overall pattern of payments, and this emerges from examining relative payments across years. Specifically, we seek to demonstrate that relative payments increase as relative disagreement with the US increases, conditional on payments being made. The example below illustrates this premise and our identification strategy.

\begin{example}[Illustration]
	Consider the set of 3 states listed in Table \ref{tab:ex1_char} in ascending order of their propensity to vote against the US (\emph{PVAUS}). We assign dollar amounts that the states require to change their vote, correlated with the \emph{PVAUS}. This minimal working example serves to illustrate how we recover the underlying correlation between a state's propensity to vote against the US and the amount of aid it receives by comparing relative changes in explanatory and outcome variables. Though not as straightforward as examining levels of disagreement and aid received, this approach is necessary for causal identification.

\begin{table}[h!]
	\vspace{10pt}
	\caption{State Characteristics for the Example}
	\label{tab:ex1_char}
	\begin{center}
		\begin{tabular}{C{0.1\textwidth}|C{0.2\textwidth}|C{0.2\textwidth}}
			State ID & PVAUS & Amount required to change vote (\$) \\
	\hline
	\hline
			1 & 1/4 & 4 \\
			2 & 1/3 & 5 \\
			3 & 1/2 & 6
	\end{tabular}
	\end{center}
\end{table}

	Suppose that in year 1, states 1 and 2 are on the council, and in year 2, state 1 is replaced by state 3. From the US's perspective, state 3's vote is more costly than state 1's. State 2 thus becomes relatively less expensive from year 1 to year 2. The cost of state 2's vote is reflected in absolute levels, but also in the relative shares, the latter of which change by way of exogenous variation in the council membership. When state 2's \textit{Contribution to Disagreement} falls in year 2, its share of the aid given to members of the council falls as well.

\begin{table}[h!]
	\vspace{10pt}
	\caption{Rotating Council Membership for Example 1}
	\label{tab:ex1_pay}
	\begin{center}
		\begin{tabular}{C{0.05\textwidth}|C{0.1\textwidth}|C{0.2\textwidth}|C{0.2\textwidth}}
			Year & State ID & \textit{Contribution to Disagreement} & Share of Aid Given \\
			\hline\hline
			1 & 1 & 3/7 & 4/9 \\
			1 & 2 & 4/7 & 5/9 \\
			\hline
			2 & 2 & 2/5 & 5/11 \\
			2 & 3 & 3/5 & 6/11 \\
		\end{tabular}
	\end{center}
\end{table}
\end{example}
\begin{center}
* * *
\end{center}

The US may not seek unanimity, and as such, states may move in and out of the set of cheapest votes required for passage of resolutions on the Council, e.g., the fifth to the seventh cheapest vote. If vote buying dies off for those countries most averse to the positions of the US, the non-monotonicity will hinder the observation of any vote buying that is occurring. To allow for this possibility, we conduct analyses both of the full sample and those states whose vote would be cheapest votes to buy, yet necessary for passage of resolutions. Indeed, this minimal-winning coalition variant of the vote buying hypotheses constitutes a more refined version of \possessivecite{Vreeland2014} prediction that the US would not reward its closest friends nor its most strident foes, but rather those potentially persuadable countries that are moderately opposed to US positions. Evidence that the pattern of payments reflects vote buying over the six members most likely to vote with the US -- but not all members -- would be evidence both of vote buying and that the US seeks only a minimal winning coalition.

\section*{Data}

As our dependent variable, we distinguish between flows of both military and economic aid. Military and economic aid are both fungible from the perspective of the recipient \cite{khilji1994fungibility,feyzioglu1998panel}, but from the perspective of the donor they may be quite different. From the perspective of the US government, it may be more palatable politically to give economic aid to ideologically distant states. However, there is evidence that, as an instrument of national security, military assistance is more clearly under the purview of the executive \cite{milner2010political}. In examining both forms of aid, we can consider such cross-cutting implications.

In some cases, a state may receive neither military nor economic aid during both years of a term on the UNSC, which indicates their lack of necessity for aid, their unwillingness to be bought, or the US's unwillingness to provide them aid. Examinations of the states receiving no aid of any sort during their UNSC term confirms that they are precisely those states most or least inclined to vote with the US as well as wealthier states. This specific sort of targeting and exclusion provides us with a censored dependent variable \cite{berthelemy2004bilateral}. As such, we follow the suggestion in the literature in using a Tobit model, which estimates the endogenous selection of aid and therefore allows us to model the data generating process that accounts for the presence of zeroes.\footnote{Tobit models have the drawback of not being consistent under fixed effects \cite{Honore1992,Wooldridge2010}. \citeasnoun{berthelemy2004bilateral} account for this by performing Tobit analysis using random effects. However, \citeasnoun{greene2004fixed} has demonstrated that the incidental parameters problem for Tobit models is not particularly grave, as the coefficients are not biased and the standard errors are only minimally biased even with a relatively short panel, particularly when compared to a random effects model in which the unit specific effects are correlated with the independent variables in the model.} %Further, \citeasnoun{berthelemy2006bilateral} has demonstrated that such foreign aid analysis is generally robust to a number of model specifications.

Because of the left-skew of the data and the long, sparsely populated right tail, it is important to log-transform the US foreign aid data. Outliers have proven to be a problem in foreign aid data \cite{burnside2000aid} and economic data more generally \cite{choi2009effect}. We show the distribution of the data before and after this transformation in Appendix \ref{sec:dists}. The transformation dramatically reduces the number of outliers in the data. Given the presence of a substantial number of zeroes in the aid data, we use the inverse hyperbolic sine function (IHS), which is a less \textit{ad hoc} approach to retaining zeroes when taking logs than adding one to each observation \cite{burbidge1988alternative}.\footnote{Letting $y$ denote aid, the inverse hyperbolic sine of $y$ is given by $\ln \left(y+\sqrt{(y^2+1)}\right)$.} Using logged data allows us to analyze proportional changes in aid, which aligns with our focus on relative changes in a state's disagreement with the US.

The Congressional budgeting process for fiscal year $ t $ occurs throughout the first two-thirds of the previous calendar year, where fiscal year $ t $ runs from October of calendar year $ t-1 $ through September of calendar year $ t $. The UNSC usually convenes during the latter quarter of calendar years, i.e., the first quarter of fiscal years. Congress would budget in the first part of calendar year $ t-1 $ the aid to be distributed for UNSC votes taken at the start of fiscal year $ t $. Hence, we match foreign aid from fiscal year $ t $ to explanatory variables reflecting calendar year $ t-1 $.

We estimate each model with and without a number of control variables. While these factors should not threaten our ability to draw causal inferences given our design, each factor may influence the willingness of the United States to grant aid for reasons other than vote buying, and thus we include these variables to reduce residual noise. First, aid from the United States might vary according to where the foreign government falls along the political spectrum, and thus we control for each state's Polity score \cite{marshallandjaggers2001}. Government turnover in the rotating member state may also influence the level of aid that the US delivers. To measure domestic political changes, we use the \emph{Change in Source of Leader Support} (CHISOLS) data set \cite{leedsandmattes2013}, which records whether or not a new leader's ascent to power is associated with a change in the underlying base of domestic support. The occurrence of a military conflict between a rotating member and the US may also curtail the amount of aid the US is willing to send. We therefore control for the onset of a militarized interstate dispute (MID) with the United States for each year using the Correlates of War (COW) Militarized Interstate Disputes dataset \cite{Jones1996}. If a potential recipient forms an alliance with the United States, the terms of the agreement may lead to a higher baseline propensity for aid. We consider whether a rotating member has a defensive or offensive alliance with the US during a given year using the Alliance Treaty Obligations and Provisions (ATOP) dataset \cite{Leeds2002}. This variable estimates only for the full sample, as there are no states that join or leave a US alliance during their UNSC term that are within the six closest member states. Lastly, we control for whether the rotating member is in its first or second year on the UN Security Council, since previous scholars demonstrated that the benefits of UNSC membership are strongest in the second year \cite{Dreher2009}.

\section*{Results}

We begin with a descriptive overview, demonstrating how many of the states in our sample display changes in the share of aid they receive that aligns with our theoretical predictions based on the change they experienced in \textit{Contribution to Disagreement}. Tables dividing our entire sample into those whose \textit{Contribution to Disagreement} decreased and increased appear in Appendix Section \ref{sec:desc_tables}. Two-thirds of the sample demonstrates behavior consistent with theory, as summarized below.

For non-permanent UNSC member states whose \textit{Contribution to Disagreement} decreases from the first to the second year of their term, vote-buying theory predicts that they receive a relatively lower share of payments made to UNSC rotating members or that, having been relatively more expensive than other votes, they received zero in the first (and possibly second) year of their term. In our sample, 70 states become relatively less expensive. With respect to military aid, 23 receive relatively more assistance in their second year, and 22 receive no assistance in their second year, meaning 64\% of the sample comports with theory. With respect to economic aid, 24 receive relatively more assistance in their second year, and 18 receive no assistance in their second year, constituting 60\% of the sample that receives payments following the predictions of vote-buying theory.

For non-permanent UNSC member states whose \textit{Contribution to Disagreement} increases from the first to the second year of their term, vote-buying theory predicts that they receive a relatively higher share of payments made to UNSC rotating members or that, having become too expensive relative to other votes, they receive zero in the second (and possibly first) year of their term. In our sample, 125 states become relatively more expensive. With respect to military aid, 51 receive relatively more assistance in their second year, and 35 receive no assistance in their second year, meaning just under 70\% of the sample comports with theory. With respect to economic aid, 55 receive relatively more assistance in their second year, and 25 receive no assistance in their second year, such that 64\% of the sample receives payments that follow the predictions of vote-buying theory.

Turning to our empirical analysis, if vote buying is occurring, we are most confident that the relationship between propensity to vote against the US and US aid will manifest itself over the six states contributing least to the UNSC's propensity to vote against the US. If the relationship holds over the entire UNSC, a desire for unanimity leads to the persuasion of all rotating members rather than just those necessary for passage of a US-supported resolution. Accordingly, after examining the six states whose propensity to vote with the US is highest, we extend the analysis to all rotating members. Our analysis covers the years 1966-2006,\footnote{This is the period in which the UNSC had 10 non-permanent members and for which we have data for our control variables.} with our unit of analysis as the state-year. For analyses of just the six states most likely to vote with the US in a given year, we accordingly lose 40\% of our observations. This leaves us with nearly 240 observations for these regressions and roughly 400 observations for regressions including all UNSC rotating members.

\begin{table}
\begin{center}
\captionsetup{justification=centering}
\caption{Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US}
\label{tab:mil}
\input{mili_aid.tex}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{center}
\end{table}

\begin{table}
\begin{center}
\captionsetup{justification=centering}
\caption{Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Economic Aid Received from the US}
\label{tab:econ}
\input{econ_aid.tex}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{center}
\end{table}

Tables \ref{tab:mil} and  \ref{tab:econ} present our main results. Columns 1 and 2 include only the members with the six lowest propensities to vote against the US  in a given year (``Rank $\leq$ 6'') -- those most easily bought. We observe a statistically and substantively significant positive relationship between \textit{Contribution to Disagreement} and both military and economic aid received. Of the members most amenable to voting with the US that constitute a near-minimal winning coalition, those relatively more prone to vote against the US receive a higher proportion of the US military and economic assistance budget doled out to UNSC members. A shift in \textit{Contribution to Disagreement} equal to one standard deviation of the within group variance (0.009) results in a 77.68\% change in military aid and an 82.56\% change in economic aid. For the median state (receiving \$558,953 in military aid), a 77.68\% increase in military aid amounts to a \$434,194 increase in military aid. Meanwhile, for the median state (a state receiving \$24,900,000 in economic aid), an 82.56\% increase in economic aid amounts to a \$20,557,440 increase in economic aid. These are substantial dollar amounts, particularly since much of this aid goes to lower and lower-middle income countries. It should be noted, however, that these estimates include states whose receipt of aid goes from zero to some strictly positive quantity. The estimates reduce somewhat in size when restricting the sample to only those states always receiving aid, though the patterns of statistical significance remain.

Columns 3 and 4 include all rotating members. \textit{Contribution to Disagreement} does not predict the outlay of military or economic aid in the full sample. This suggests that the US may be unwilling to provide assistance to those states that display the least affinity towards it. Thus, evidence that UNSC decisions are often unanimous does not result from vote buying, but rather a combination of the true voting preferences of states and seeing little value in going against the majority.\footnote{The counterintuitive finding of a positive and significant coefficient for a military dispute with the US in column 2 of the military assistance regressions is driven by a single observation: Panama in 1976. In this dispute, a Panamanian gunboat held two US vessels in response to illegal fishing. This was a minor altercation in a time of significant diplomatic negotiations regarding future control over and defense of the Panama Canal, for which the US increased its aid to Panama. Our results are robust to the exclusion of this variable, though it is theoretically important to include, and the variable only achieves significance in this single regression.}

In Appendix \ref{sec:rob_plac}, we present robustness checks, placebo tests, and extensions of our baseline analyses. First, our results are not highly sensitive to allowing for slightly greater than the six states least likely to vote against the US using military aid, but they are when examining economic aid. This result is somewhat surprising in light of the belief that delivering military aid to a highly disparate state would be more costly for the leader. This could potentially reflect uncertainty over individual votes. Further, the effect weakens but remains significant for both military and economic aid if we examine only the five closest states, suggesting that the US may indeed seek to account for the possibility of an abstention by Russia or China, though not in every instance. Tables \ref{tab:othercut_mil_boot} and \ref{tab:othercut_econ_boot} present these findings.

Propensity to disagree with the United States is, of course, a proxy for how amenable a state would be to selling its vote to the US. Other factors that may influence how expensive it is to buy a state's vote include its economic health. As such, in Tables \ref{tab:mili_gdp} and \ref{tab:econ_gdp}, we measure US aid as a percentage of a state's GDP as our dependent variable. The pattern in the baseline results holds in these models, demonstrating that our results are not an artifact of other factors behind the cost of a country's vote besides its likelihood to vote with the US.

As a placebo test in Tables \ref{tab:lp_mil} and \ref{tab:lp_econ}, we match aid from year $t+2$ to explanatory variables from year $t$.  Our identification holds only for within-term changes in relative disagreement with the US, so the aid distributed after a state's UNSC term is unrelated to our predictions. We are testing for false positives, and the null results provide evidence that our main results are neither spurious nor do they arise from some factor outside the UNSC rotation. This test therefore serves a different purpose than the \citeasnoun{Kuziemko2006} test that finds that aid returns essentially to the pre-election baseline when states exit the council. While some tests regarding economic aid display statistical significance, the estimates are in the opposite direction of what vote buying theory would predict and implausibly large. Not finding evidence in favor of vote buying in this context allays concerns that our results are unrelated to the proposed mechanism.

We also consider the possibility that the strategic nature of aid has changed over time. In particular, the foreign aid literature has suggested aid has become less geopolitical and more specifically conditional since the end of the Cold War \cite{dunning2004conditioning,bermeo2011foreign}. We therefore examine the allocation of military aid in the Cold War period and the post-Cold War period. Tables \ref{tab:mili_aid_cold} and \ref{tab:mili_aid_pcold} present these findings. We find consistent and similar effects for the Cold War period. Although we see the effect trends in the same direction for the six closest states, we do not find significant results for the post-Cold War period. This would be consistent with the previous literature. However, the post-Cold War models reduce our sample by nearly two-thirds. Since we are leveraging only a small amount of variation to guard against possible problems of endogeneity, this is a substantial decrease, as the standard errors for the model demonstrate. Thus, it is difficult to fully evaluate whether this result is driven by post-Cold War dynamics or a lack of sufficient data. Further following this literature, we find some evidence of disparate results when separately examining democratic and non-democratic states in Tables \ref{tab:mili_demvaut} and \ref{tab:econ_demvaut}. The US appears to be more willing to use foreign aid to to buy votes on the UNSC from non-democratic states.

Closely following the predictions of vote buying theory, we find substantial evidence that the United States uses foreign aid to buy votes from rotating members on the UNSC. While the effects we uncover are large, we note that they are similar to effects reported in the related analyses of \citeasnoun{Kuziemko2006} and \citeasnoun{Vreeland2014}, while overcoming some of the limitations in this previous work.

\section*{Discussion}

Previous studies establish a connection between membership on the United Nations Security Council and receipt of foreign assistance, particularly from the United States. These studies argued that this pattern suggested attempts by the US to buy influence, but they stop short of supplying conclusive causal analyses of this claim. We substantiate this claim by presenting causally-identified evidence that the allocation of military and economic aid from the US to UNSC members reflects a core prediction of vote buying theory. We observe a statistically and substantively significant positive relationship between a rotating member's relative propensity to vote against the US and military and economic aid received from the US in procurement of a minimal winning coalition. We see little evidence that the US disperses aid to persuade the rest of the UNSC rotating members in service of procuring unanimity.

Our findings also constitute evidence that patterns of political behavior present in domestic politics are present at the level of international institutions. Future scholarly work might examine the role the other permanent members play in vote buying on the UNSC. This is especially interesting in light of the increasing role of China as both an emerging donor of foreign aid \cite{woods2008whose} and as a global counterweight to the United States, as well as President Donald Trump's proposed cuts to the foreign aid budget. Further, this methodology can be used to analyze whether other forms of international aid, such as IMF loans and World Bank development aid, constitute attempts at vote buying.

The effect identified above entails substantial sums of money. The finding that a state's receipt of bilateral aid from the United States may increase by over 75\% relative to what it may have otherwise expected reflects a substantial shock to that state's resources.\footnote{This increase is above and beyond the increase due to serving on the UNSC.} Yet this is perhaps the most trivial of the ways in which we may gauge the role of vote buying. Perhaps most importantly, these findings call into question how we should view UNSC decisions.

As discussed, collective legitimization is one the major political functions of the United Nations more generally \cite{Claude1966} and specifically the UNSC \cite{hurd2002legitimacy}. Security Council decisions often dictate the global system's political approval or disapproval regarding the policies and actions of states. If the US is capable of buying votes on the UNSC, this calls into question the United Nations as an independent actor capable of exercising collective authority for the international community. Evidence of vote buying undermines the success of multilateral governance more generally, as this form of multilateralism does not reflect multilateral decision making.\footnote{Indeed, even in the UNSC's most recent strong rebuke of US foreign policy, their refusal to support the US invasion of Iraq, it was the permanent members of the UNSC, rather than the rotating members, that prevented a resolution backing the use of military force.} While scholars had previously noted the minimal voting power of temporary members \cite{o1996power}, we find that these members are willing to sell what little power they have to the dominant states in the system. This is particularly alarming evidence in light of the finding that despite this increased foreign assistance, the rotating members on the UNSC have lower levels of economic growth and democracy following their time on the Security Council \cite{bueno2010pernicious}. Thus while their vote is important for the legitimacy of international action, temporary members are both unable to exert political influence in the international community and unable to leverage their vote towards improving the health of their nation.

\clearpage
\section*{Supplemental Information}

Online appendices and replication materials can be found at http://d-alexander.com, https://bryanandrewrooney.wordpress.com, and at the International Studies Quarterly data archive.

\clearpage
\singlespacing
\bibliographystyle{apsr}
\bibliography{library.bib}

%\begin{comment}

\appendix
\clearpage
\section{Construction of the Explanatory Variable}
\label{sec:const}
Our measure of states' propensity to vote against the US, which we call the Probability of Voting Against the US (PVAUS), results from the following regression. The dependent variable takes the value of 0 if a state voted the same way as the US, 1 if the state voted the opposite as the US, and 0.5 if a state abstained but the US cast a yes or no vote. We regress this outcome, for all combinations of states, $ i $, and General Assembly resolutions, $ j $, in a given year, $ t $, on which the US did not abstain, on state-specific fixed effects. Our PVAUS measure is simply the estimated coefficients for the state fixed effects, as this represents the state's latent propensity to vote against the US.%***
\begin{equation}
\label{eq:PVAUS}
\mathbb{I}(vote_{i,j,t}\neq vote_{US,j,t}) = \textit{PVAUS}_{i,t} + \eta_{i,j,t}\footnote{In fact, the LHS may take values of 0.5 in the case of state $i$'s abstention on resolution $j$. Note also, we suppress the constant such that PVAUS takes values between 0 and 1.}
\end{equation}
As discussed, if state $ i $'s two-year term on the UNSC consists of years $ t $ ($ SC_1 $) and $ t+1 $ ($ SC_2 $), its USVP from the year preceding $ i $'s term, $ t-1 $, comprises the fixed, lagged PVAUS, which is denoted by $ \overline{\text{PVAUS}} $.
\begin{equation}
\label{eq:FIXLAG}
\overline{\text{PVAUS}}:= \text{PVAUS}_{SC_1,SC_2} = \text{PVAUS}_{t-1}
\end{equation}
We then obtain the relative propensity of a state on the UNSC to vote against the US in a given year based on their PVAUS relative to the PVAUS of all other members of the UNSC, as specified in equation \ref{eq:cont}. We refer to this as a state's \textit{Contribution to DIsagreement}, as discussed in text.
\begin{equation}
\label{eq:cont}
\textit{Contribution to Disagreement}_{i,t} = \text{PVAUS}_{i,t}\left/\sum_{k\text{ on UNSC in year }t} \text{PVAUS}_{k,t}\right.
\end{equation}
and perform the regression analysis specified by equation \ref{eq:SPEC},
\begin{equation}
\label{eq:SPEC}
\ln(AID_{i,t+1}) = \alpha_i + \gamma_t + \varphi_{i,t}\textit{Contribution to Disagreement}_{i,t} + \mathbf{X}_{i,t}\beta + \epsilon_{i,t}.
\end{equation}

\clearpage
\section{Distribution of the Dependent Variable}
\label{sec:dists}

\begin{figure}[!ht]
\centering
\begin{minipage}{.5\textwidth}
  \centering
  \includegraphics[width=1\linewidth]{econ_aid.pdf}
  \label{foreign_econ}
\end{minipage}%
\begin{minipage}{.5\textwidth}
  \centering
  \includegraphics[width=1\linewidth]{ihs_econ.pdf}
  \label{domestic_econ}
\end{minipage}
\end{figure}

\begin{figure}[!ht]
\centering
\begin{minipage}{.5\textwidth}
  \centering
  \includegraphics[width=1\linewidth]{mil_aid.pdf}
  \label{foreign_mili}
\end{minipage}%
\begin{minipage}{.5\textwidth}
  \centering
  \includegraphics[width=1\linewidth]{ihs_mil.pdf}
  \label{domestic_mili}
\end{minipage}
\end{figure}

\clearpage
\section{Descriptive Analysis of the Data}
\label{sec:desc_tables}

\begin{table}[!h]
\caption{States Whose \textit{Contribution to Disagreement} Decreased}
\vspace{12pt}
\resizebox{0.5\textwidth}{!}{
\begin{tabular}{L{.225\textwidth}|C{.075\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}}
Non-Permanent UNSC Member State	&	Second Year	&	$\downarrow$ Military Aid	&	Zero Military Aid in First Year	&	$\downarrow$ Economic Aid	&	Zero Economic Aid in First Year	\\
\hline\hline
Algeria	&	2005	&		&		&		&		\\
Argentina	&	1972	&	x	&		&		&	x	\\
Argentina	&	1995	&		&		&	x	&		\\
Austria	&	1992	&	x	&		&		&		\\
Belgium	&	1972	&	x	&		&		&	x	\\
Belgium	&	1992	&	x	&		&		&		\\
Benin	&	2005	&		&	x	&		&		\\
Bolivia	&	1979	&	x	&		&	x	&		\\
Brazil	&	1994	&		&		&	x	&		\\
Brazil	&	2005	&		&		&		&		\\
Bulgaria	&	1987	&		&	x	&		&	x	\\
Burundi	&	1971	&		&	x	&		&		\\
Cabo Verde	&	1993	&		&	x	&		&	x	\\
Canada	&	1978	&		&	x	&		&	x	\\
Chile	&	1997	&		&		&		&		\\
Colombia	&	1970	&		&		&		&		\\
Congo	&	1987	&	x	&		&	x	&		\\
Costa Rica	&	1998	&		&		&		&		\\
Cuba	&	1991	&		&	x	&	x	&		\\
Czech Republic	&	1995	&	x	&		&	x	&		\\
Czechoslovakia	&	1979	&	x	&		&	x	&		\\
Dem. Rep. of Congo	&	1991	&	x	&		&	x	&		\\
Djibouti	&	1994	&		&		&		&		\\
Ecuador	&	1992	&	x	&		&		&		\\
Egypt	&	1997	&	x	&		&		&		\\
Finland	&	1970	&		&		&		&	x	\\
Gabon	&	1979	&		&	x	&	x	&		\\
Germany	&	1978	&		&	x	&		&	x	\\
Ghana	&	1987	&		&		&	x	&		\\
Guinea-Bissau	&	1997	&	x	&		&	x	&		\\
Hungary	&	1993	&		&		&		&		\\
India	&	1978	&		&		&		&		\\
India	&	1992	&	x	&		&		&		\\
Italy	&	1972	&		&	x	&		&	x	\\
Ivory Coast	&	1991	&		&	x	&	x	&		\\
\hline
\end{tabular}
}
\resizebox{0.5\textwidth}{!}{
\begin{tabular}{L{.225\textwidth}|C{.075\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}}
Non-Permanent UNSC Member State	&	Second Year	&	$\downarrow$ Military Aid	&	Zero Military Aid in First Year	&	$\downarrow$ Economic Aid	&	Zero Economic Aid in First Year	\\
\hline\hline
Japan	&	1972	&	x	&		&	x	&		\\
Japan	&	1993	&		&	x	&		&	x	\\
Japan	&	1998	&		&	x	&		&	x	\\
Kenya	&	1998	&		&		&		&		\\
Kuwait	&	1979	&	x	&		&	x	&		\\
Mauritius	&	1978	&		&	x	&		&		\\
Morocco	&	1993	&		&		&		&		\\
Nepal	&	1970	&		&		&		&		\\
New Zealand	&	1994	&		&	x	&	x	&		\\
Nicaragua	&	1971	&		&		&		&		\\
Nigeria	&	1979	&	x	&		&	x	&		\\
Nigeria	&	1995	&	x	&		&	x	&		\\
Oman	&	1995	&		&		&	x	&		\\
Pakistan	&	1994	&		&	x	&	x	&		\\
Philippines	&	2005	&		&		&		&		\\
Poland	&	1971	&		&	x	&		&	x	\\
Poland	&	1997	&		&		&		&		\\
Portugal	&	1998	&		&		&		&	x	\\
Rep. of Korea	&	1997	&	x	&		&		&	x	\\
Romania	&	1991	&		&	x	&	x	&		\\
Romania	&	2005	&		&		&		&		\\
Rwanda	&	1995	&	x	&		&	x	&		\\
Sierra Leone	&	1971	&		&	x	&		&		\\
Somalia	&	1972	&	x	&		&		&	x	\\
Spain	&	1970	&		&		&		&	x	\\
Spain	&	1994	&		&		&	x	&		\\
Sweden	&	1998	&		&	x	&		&	x	\\
Syria	&	1971	&		&	x	&		&		\\
UAE	&	1987	&		&	x	&		&	x	\\
Venezuela	&	1978	&	x	&		&		&	x	\\
Venezuela	&	1987	&		&		&		&		\\
Venezuela	&	1993	&		&		&		&		\\
Yemen	&	1991	&	x	&		&	x	&		\\
Zambia	&	1970	&		&	x	&	x	&		\\
Zimbabwe	&	1992	&	x	&		&		&		\\
\hline
\end{tabular}
}

\vspace{6pt}
\footnotesize{\textit{Notes:}
For non-permanent UNSC member states whose \textit{Contribution to Disagreement} decreases from the first to the second year of their term, vote-buying theory predicts that they receive a relatively lower share of payments made to UNSC rotating members or that, having been relatively more expensive than other votes, they received zero in the first (and possibly second) year of their term. In our sample, 70 states become relatively less expensive. With respect to military aid, 23 receive relatively more assistance in their second year, and 22 receive no assistance in their second year, meaning 64\% of the sample comports with theory. With respect to economic aid, 24 receive relatively more assistance in their second year, and 18 receive no assistance in their second year, constituting 60\% of the sample that receives payments following the predictions of vote-buying theory.}
\end{table}

\clearpage
\begin{table}
\caption{States Whose \textit{Contribution to Disagreement} Increased}
\vspace{12pt}
\resizebox{0.5\textwidth}{!}{
\begin{tabular}{L{.225\textwidth}|C{.075\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}}
Non-Permanent UNSC Member State	&	Second Year of Term	&	$\uparrow$ Military Aid	&	Zero Military Aid in Second Year	&	$\uparrow$ Economic Aid	&	Zero Economic Aid in Second Year	\\
\hline\hline
Algeria	&	1969	&		&	x	&		&		\\
Algeria	&	1989	&		&		&		&	x	\\
Angola	&	2004	&	x	&		&		&		\\
Argentina	&	1988	&		&		&	x	&		\\
Argentina	&	2000	&		&		&		&		\\
Argentina	&	2006	&	x	&		&	x	&		\\
Australia	&	1974	&	x	&		&	x	&		\\
Australia	&	1986	&	x	&		&	x	&		\\
Austria	&	1974	&	x	&		&	x	&		\\
Bahrain	&	1999	&		&		&	x	&		\\
Bangladesh	&	1980	&		&		&		&		\\
Bangladesh	&	2001	&	x	&		&	x	&		\\
Belarus	&	1975	&	x	&		&	x	&		\\
Benin	&	1977	&		&	x	&	x	&		\\
Botswana	&	1996	&		&		&		&		\\
Brazil	&	1968	&		&		&		&		\\
Brazil	&	1989	&		&		&		&		\\
Brazil	&	1999	&		&		&		&		\\
Bulgaria	&	2003	&		&		&		&		\\
Burkina Faso	&	1985	&	x	&		&	x	&		\\
Cameroon	&	1975	&	x	&		&	x	&		\\
Cameroon	&	2003	&		&		&		&		\\
Canada	&	1968	&		&	x	&		&	x	\\
Canada	&	1990	&	x	&		&		&	x	\\
Canada	&	2000	&		&		&		&	x	\\
Chile	&	2004	&	x	&		&		&		\\
Colombia	&	1990	&	x	&		&		&		\\
Colombia	&	2002	&	x	&		&		&		\\
Costa Rica	&	1975	&	x	&		&	x	&		\\
Dem. Rep. of Congo	&	1983	&	x	&		&	x	&		\\
Denmark	&	1968	&		&	x	&		&	x	\\
Denmark	&	1986	&	x	&		&	x	&		\\
Denmark	&	2006	&	x	&		&	x	&		\\
Egypt	&	1985	&	x	&		&	x	&		\\
Ethiopia	&	1968	&		&		&		&		\\
Ethiopia	&	1990	&		&	x	&		&		\\
Finland	&	1990	&	x	&		&		&	x	\\
GDR	&	1981	&		&	x	&	x	&		\\
Gabon	&	1999	&	x	&		&		&		\\
Gambia	&	1999	&		&	x	&		&	x	\\
Germany	&	1988	&		&	x	&		&	x	\\
Germany	&	1996	&		&	x	&		&	x	\\
Germany	&	2004	&		&		&		&		\\
Greece	&	2006	&		&		&	x	&		\\
Guinea	&	1973	&		&	x	&		&		\\
Guinea	&	2003	&		&		&		&		\\
Guyana	&	1976	&		&	x	&		&		\\
Guyana	&	1983	&		&	x	&		&		\\
Honduras	&	1996	&		&		&		&		\\
Hungary	&	1969	&		&	x	&	x	&		\\
India	&	1968	&		&		&		&		\\
India	&	1973	&	x	&		&		&		\\
India	&	1985	&	x	&		&	x	&		\\
Indonesia	&	1974	&	x	&		&	x	&		\\
Indonesia	&	1996	&		&		&		&		\\
Iraq	&	1975	&	x	&		&	x	&		\\
Ireland	&	1982	&	x	&		&		&	x	\\
Ireland	&	2002	&		&		&	x	&		\\
Italy	&	1976	&		&	x	&		&		\\
Italy	&	1988	&		&	x	&		&	x	\\
Italy	&	1996	&		&	x	&		&	x	\\
Jamaica	&	1980	&	x	&		&		&		\\
Jamaica	&	2001	&	x	&		&	x	&		\\
\hline
\end{tabular}
}
\resizebox{0.5\textwidth}{!}{
\begin{tabular}{L{.225\textwidth}|C{.075\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}|C{.1\textwidth}}
Non-Permanent UNSC Member State	&	Second Year of Term	&	$\uparrow$ Military Aid	&	Zero Military Aid in Second Year	&	$\uparrow$ Economic Aid	&	Zero Economic Aid in Second Year	\\
\hline\hline
Japan	&	1976	&		&	x	&		&	x	\\
Japan	&	1982	&	x	&		&		&	x	\\
Japan	&	1988	&		&	x	&		&	x	\\
Japan	&	2006	&	x	&		&	x	&		\\
Jordan	&	1983	&	x	&		&	x	&		\\
Kenya	&	1974	&	x	&		&	x	&		\\
Libya	&	1977	&		&	x	&	x	&		\\
Madagascar	&	1986	&	x	&		&	x	&		\\
Malaysia	&	1990	&		&	x	&		&		\\
Malaysia	&	2000	&		&		&		&		\\
Mali	&	2001	&	x	&		&	x	&		\\
Malta	&	1984	&		&	x	&	x	&		\\
Mauritania	&	1975	&	x	&		&	x	&		\\
Mauritius	&	2002	&	x	&		&		&		\\
Mexico	&	1981	&		&		&	x	&		\\
Mexico	&	2003	&		&		&	x	&		\\
Namibia	&	2000	&		&		&		&		\\
Nepal	&	1989	&		&		&		&		\\
Netherlands	&	1984	&		&	x	&	x	&		\\
Netherlands	&	2000	&		&	x	&		&	x	\\
Nicaragua	&	1984	&		&	x	&		&	x	\\
Niger	&	1981	&		&		&	x	&		\\
Norway	&	1980	&		&	x	&		&	x	\\
Norway	&	2002	&	x	&		&		&	x	\\
Pakistan	&	1969	&		&		&		&		\\
Pakistan	&	1977	&		&		&	x	&		\\
Pakistan	&	1984	&		&		&	x	&		\\
Pakistan	&	2004	&	x	&		&		&		\\
Panama	&	1973	&		&		&		&		\\
Panama	&	1977	&		&		&	x	&		\\
Panama	&	1982	&	x	&		&		&		\\
Paraguay	&	1969	&		&		&		&		\\
Peru	&	1974	&	x	&		&	x	&		\\
Peru	&	1985	&		&		&	x	&		\\
Philippines	&	1981	&		&		&	x	&		\\
Poland	&	1983	&	x	&		&		&		\\
Portugal	&	1980	&		&		&		&		\\
Romania	&	1977	&		&	x	&	x	&		\\
Senegal	&	1969	&		&		&		&		\\
Senegal	&	1989	&		&		&		&		\\
Singapore	&	2002	&	x	&		&		&	x	\\
Slovenia	&	1999	&		&		&		&	x	\\
Spain	&	1982	&	x	&		&		&		\\
Spain	&	2004	&		&	x	&	x	&		\\
Sudan	&	1973	&		&	x	&		&	x	\\
Sweden	&	1976	&		&	x	&		&	x	\\
Syria	&	2003	&		&	x	&	x	&		\\
Tanzania	&	1976	&		&	x	&		&		\\
Tanzania	&	2006	&	x	&		&	x	&		\\
Thailand	&	1986	&		&		&	x	&		\\
Togo	&	1983	&	x	&		&	x	&		\\
Trinidad and Tobago	&	1986	&	x	&		&	x	&		\\
Tunisia	&	1981	&	x	&		&	x	&		\\
Tunisia	&	2001	&	x	&		&	x	&		\\
Uganda	&	1982	&	x	&		&		&		\\
Ukraine	&	1985	&	x	&		&	x	&		\\
Ukraine	&	2001	&	x	&		&	x	&		\\
Yugoslavia	&	1973	&		&	x	&		&	x	\\
Yugoslavia	&	1989	&		&	x	&		&	x	\\
Zambia	&	1980	&		&	x	&		&		\\
Zambia	&	1988	&		&	x	&		&		\\
Zimbabwe	&	1984	&		&		&	x	&		\\
\hline
\end{tabular}
}

\vspace{6pt}
\footnotesize{\textit{Notes:}
For non-permanent UNSC member states whose \textit{Contribution to Disagreement} increases from the first to the second year of their term, vote-buying theory predicts that they receive a relatively higher share of payments made to UNSC rotating members or that, having become too expensive relative to other votes, they receive zero in the second (and possibly first) year of their term. In our sample, 125 states become relatively more expensive. With respect to military aid, 51 receive relatively more assistance in their second year, and 35 receive no assistance in their second year, meaning just under 70\% of the sample comports with theory. With respect to economic aid, 55 receive relatively more assistance in their second year, and 25 receive no assistance in their second year, such that 64\% of the sample receives payments that follow the predictions of vote-buying theory.}
\end{table}

\clearpage
\section{Robustness Checks, Placebo Tests \& Extensions}
\label{sec:rob_plac}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Robustness Check: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US with Varying Rank Cutpoints}
\label{tab:othercut_mil_boot}
\input{cut_mili_aid.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\clearpage
\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Robustness Check: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Economic Aid Received from the US with Varying Rank Cutpoints}
\label{tab:othercut_econ_boot}
\input{cut_econ_aid.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\clearpage
\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Robustness Check: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US as a \% of GDP}
\label{tab:mili_gdp}
\input{mili_aid_gdp.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Robustness Check: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Economic Aid Received from the US as a \% of GDP}
\label{tab:econ_gdp}
\input{econ_aid_gdp.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\clearpage
\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Placebo Test: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US During Following Year}
\label{tab:lp_mil}
\input{placebo_mili_aid.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Placebo Test: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Economic Aid Received from the US During Following Year}
\label{tab:lp_econ}
\input{placebo_econ_aid.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Extension: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US During the Cold War}
\label{tab:mili_aid_cold}
\input{mili_aid_cold.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\begin{comment}
\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Extension: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Economic Aid Received from the US During the Cold War}
\label{tab:econ_aid_cold}
\input{econ_aid_cold.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}
\end{comment}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Extension: Analyses of the Effect of \textit{Contribution to Disagreement} with the US on Military Aid Received from the US After the Cold War}
\label{tab:mili_aid_pcold}
\input{mili_aid_pcold.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\clearpage
\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Extension: Analyses of the Effect of \textit{Contribution to Disagreement} with the US and Democracy on Military Aid Received from the US}
\label{tab:mili_demvaut}
\input{mili_aid_demvaut.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}

\begin{table}[!ht]
\begin{center}
\captionsetup{justification=centering}
\caption{Extension: Analyses of the Effect of \textit{Contribution to Disagreement} with the US and Democracy on Economic Aid Received from the US}\label{tab:econ_demvaut}
\input{econ_aid_demvaut.tex}
\end{center}
\footnotesize \textit{Notes:} Robust standard errors appear in parentheses below the estimated coefficients of all models. \\
 {*} \(p<0.10\), {**} \(p<0.05\), {***} \(p<0.01\)
\end{table}


\end{document}
